by Avi Gesser, Jehan A. Patterson, Tricia Bozyk Sherno, Frank Colleluori, and Anna R. Gressel
We recently wrote about how rights-based regulatory regimes for artificial intelligence (as opposed to risk-based frameworks) can lead to a misallocation of resources because compliance will require too much effort on low-risk AI (e.g., spam filters, graphics generation for games, inventory management, etc.) and not enough effort on AI that can actually pose a high risk of harm to consumers or the public (e.g., hiring, lending, underwriting, etc.). In this follow-up blog post, we discuss why regulators should view AI risk the same way as employee risk for large companies, and accordingly adopt risk-based regulatory frameworks for AI.